Back to Top

Paper Title

Corrigenda to Satorra, A., and Bentler, P.M. (2010), “Ensuring Positiveness of the Scaled Difference Chi-Square Test Statistic,” Psychometrika, 75, pp. 243–248

Keywords

  • Positive Definite Matrix
  • Matrix Algebra
  • Scaled Difference Chi-Square Test
  • Structural Equation Modeling (SEM)
  • Trace of Matrix
  • Model Comparison
  • Matrix Rank
  • Matrix Decomposition
  • Matrix Product
  • Statistical Robustness
  • Chi-Square Test
  • Model Evaluation
  • Mathematical Correction
  • Statistical Theory
  • Asymptotic Theory
  • Covariance Structure Analysis

Article Type

Research Article

Research Impact Tools

Issue

Volume : 75 | Page No : 243–248

Published On

February, 2025

Downloads

Abstract

On page 245, lines 3 and 4, of the published paper, we find the following text: “Since tr{UdΓ} can be expressed as the trace of the product of two positive definite matrices, tr{UdΓ} > 0, and thus cd > 0;” This text should be replaced with: “Since tr{UdΓ} can be expressed as the trace of a positive definite matrix, tr{UdΓ} > 0, and thus cd > 0;” The uncorrected text claims that Ud and Γ are positive definite matrices, but Ud can’t be positive definite, since its rank (difference between the ranks of the derivatives of the two models involved) is much less than its order. The expression tr{UdΓ} could be written differently so that the conclusion still holds. Namely, write Ud = VΠP−1 A′ (AP−1 A′ )−1 AP−1 Π′ V (formula (4) of the paper) as Ud = FF′ , where F = VΠP−1 A′ (AP−1 A′ )−1/2 ; then, tr{UdΓ} = tr{FF′ Γ} = tr{F′ ΓF}, where F′ ΓF is a positive definite matrix, given that Γ is positive definite in the setup of the paper. For rewriting the alternative expression of tr{UdΓ}, we used the well-known matrix algebra result that tr{MN} = tr{NM} for matrices M and N of dimensions conformable with the products; in our application, M = F and N = F′ Γ.

View more >>

Uploded Document Preview